Re: [wsbpel] Issue 103 - Good Idea!

[Alex Yiu <alex.yiu@oracle.com>]

Hi all,

[After a morning of meeting, back online ... :-) ]

Yes, I agree with Ron.
We may not want to overload expression, like $var/a/b/c, as the 
assignment target.
I tend to believe that we still want to:
(1) keep both expression and query in from-spec
(2) keep query in to-spec
(3) allow the usage of $var in expression

Danny ... if we pass issue 11 (assume we go for insertBefore and 
appendChild), the to-spec for  insertBefore and appendChild will be the 
same. The only change is: expression in from-spec now can use $var

E.g.:
<appendChild>
   <from expression="$fooVar/abc/def" />
   <to variable="barVar" query="/ghi/jkl" />
</appendChild>

the expression would be appended as as the last child under "/ghi/jkl". 
And, "/ghi/jkl" are existing nodes.

IMHO, XPath location path used in query or expression will always point 
to existing nodes, even after / if Issue 11 is passed.



Thanks!



Regards,
Alex yiu




Ron Ten-Hove wrote:

> Danny van der Rijn wrote:
>
>> maybe i just don't know XPATH well enough, but if $var/a/b/c doesn't 
>> exist,
>> can it be the target of an assign?
>>
>>  
>>
> XPath 1.0 expressions turn into node-sets, booleans, numbers, or 
> strings (and do so without side effects). They don't turn into 
> positional references. $var/a/b/c doesn't make sense as the target of 
> an <assign> activity, from the XPath point of view.
>
> -Ron
>
> To unsubscribe from this mailing list (and be removed from the roster 
> of the OASIS TC), go to 
> http://www.oasis-open.org/apps/org/workgroup/wsbpel/members/leave_workgroup.php. 
>
>